- Email: [email protected]
Diet and prostate cancer: a case-control study
European3oumalof
Cancervol. 33, No. 1, pp. 101-107, 1997 c: 1997 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0959.8049197 $17.00+0.00
Pergamon PII: SO959-8049(96)00373-S
Original Paper Diet and Prostate Cancer: a Case-Control H.D. Vlajinac,’ ‘Institute
of Epidemiology,
‘Institute
of Social Belgrade;
J.M. MarinkoviC,2
School
Medicine, and School
of Medicine,
Statistics
Belgrade
and Health
of Medicine,
M.D.
University,
Research,
Kragujevac
Ilit
School
University,
Study
and N.I. Kocev’ Visegradska
26,
11000,
of Medicine,
Belgrade
Kragujevac,
Yugoslavia
Belgrade; University,
A case-control study., performed in two towns of Serbia (Yugoslavia) from 1990 to 1994, comprised 101 patients with histologically confirmed prostate cancer and 202 hospital controls individually matched by age (k2 years), hospital admittance and place of residence. Dietary information was obtained by using a standard questionnaire. After adjustment for possible confounders, risk factors for prostate cancer appeared to be the highest tertile of protein (odds ratio (OR) = 13.54, 95% confidence interval (CI) = 2.38-77.13), saturated fatty acid (OR = 3.63, 95% CI = 1.03-12.79), fibre (OR = 4.02, 95% CI = 1.38-11.73), and vitamin B12 intake (OR = 2.07, 95% CI = 1.08-3.97); a protective effect was found for the highest tertile of a-tocopherol (OR = 0.15, 95% CI = 0.05-0.53), calcium (OR = 0.37, 95% CI = 0.14-0.99) and iron intake (OR = 0.34, 95% CI = 0.12-0.95). There were significant (PC 0.05) linear trends in the odds ratios for c+tocopherol, vitamin B12, calcium and iron. According to logistic regression step by step analysis, risk factors for prostate cancer were dietary intake of retinol equivalent (OR = 1.64, 95% CI = 1.01-2.67) and vitamin B12 (OR = 1.87, 95% CI = 1.15-3.05), and a protective effect was found for dietary intake of iron (OR = 0.40, 95% CI = 0.27-0.58). 0 19!#7 Published by Elsevier Science Ltd. All rights reserved. Key words: prostate cfancer, diet, case-control EurJ
Cancer,
Vol. 33, No.
1, pp. 101-107,
study
1997
INTRODUCTION LrrrLEISknown about the aetiology of prostate cancer. A
1994. Of 141 patients with histologically confirmed clinical prostate cancer, 12 patients could not be interviewed as they gave incorrect addresses, 9 patients refused to participate, 10 patients were too ill to be interviewed and 9 patients died before we were able to contact them. The final group consisted of 101 prostate cancer patients. For each case, two hospital controls (202 controls in total) were chosen among patients confirmed as having neither prostate cancer nor other prostate diseases. Those with other malignancies were also excluded. The majority of controls-154 patients, were treated at the hospital because of physical injuries, 11 had asthma, 8 had pneumonia, 8 had a peptic ulcer, 7 had cholecistitis, 6 had angina pectoris, 4 had cirrhosis, 3 had pleuritis and 1 had pancreatitis. All selected controls were interviewed-no one refused to participate. Cases and controls were individually matched by age (f2 years), hospital admittance and place of residence. Demographic, epidemiological and dietary data were obtained using a standard questionnaire. Dietary information was obtained by a quantitative history approach in which subjects were asked about their usual frequency of
number of risk factors h.ave been identified in past epidemiological studies, although the findings have not always been reproducible. The plausibility of diet having a major role in the aetiology of ptostate cancer is supported by certain observations based on descriptive epidemiology of the disease [l], but the results of analytical epidemiological studies have been inconsistent. This report describes results relating to dietary factors, obtained in a case-control study carried out in Serbia. A quantitative dietary history was used to estimate the weekly intake of different nutrients. Results related to non-dietary factors will be reported separately.
MATERIALS AND METHODS The study was conducted in two towns of Central Serbia (Kragujevac and Cuprija) from January 1990 to December Correspondence to H. Vlajinac. Received 20 Feb. 1996; revised 26 Jul. 1996; accepted 3 Sep. 1996. 101
102
H.D. Vlajinac
intake and portion size of a list of 150 food items including alcoholic beverages. The technique was similar to the one used by Jain and associates [2], although somewhat modified and adapted to suit the Serbian diet. Participants were asked how many months per year they used each food item and how often (daily/weekly/monthly/yearly). To facilitate quantification of intakes, 21 photographs of food items in three different portion sizes of known quantity were used in the interview. As measurements of consumption, standard cups, spoons, slices etc. were used. The referent period was 10 years before the disease. If significant changes in the diet had happened during the period observed, dietary habits before the change were recorded. Subjects were also questioned about their intake of vitamin and mineral supplements. From data obtained, the total daily amount of consumption for each food item was calculated and then finally converted into nutrients using food consumption tables [3]. For statistical analysis of data, univariate and multivariate logistic regression methods were used [4, 51. Tertiles of average daily intake of control group were used as a basis for comparison. A test for linear trend in risk was calculated as proposed by Breslow and Day [4]. RESULTS Table 1 describes basic demographic and personal characteristics of the case and control subjects. There were no notable differences between the groups. Table 2 gives the mean values of daily intake of nutrients. According to univariate logistic regression analysis, the mean intake of carbohydrates, sugar, fibre, tocopherol, thiamin, vitamin B6, sodium, potassium, phosphorus and iron and the mean total energy were significantly different between cases and controls. These variables, together with the mean inake of polyunsaturated fatty acids, cholesterol, vitamin B12 and magnesium, related to prostate cancer at a significant level of P < 0.10 entered into the model of multivariate analysis. According to multivariate logistic regression analysis, independent positive association with prostate cancer was found for the mean daily intake of vitamin B12 (P= 0.000) and a negative association was found for the
Table
1. Characteristics
of prostate
cancer
patients
and
their
controls
Characteristic Mean age (years) Ever married (%) Education (years) % o-4 5-12 >12 Occupation (%) Farmers Workers Clerks Place of residence (%) Urban Rural Mean height (cm) Mean weight (kg) Mean body mass index
Cases (n = 101)
Controls (I? = 202)
70.5
71.5
100.0
99.5
58.4 37.6 4.0
64.8 29.2 5.9
37.6 44.6 17.6
35.6 50.0 14.4
48 52 172.9 75.3 25.2
48 52 174.2 74.9 24.3
et al.
Table 2. Mean
values
(standard
deviation)
of daily intake
of
dietary nutrients
Nutrient
(unit)
Energy (kcal) Protein (g) Fat-total (g) Fat-animal (g) Saturated fatty acid (g) Monounsaturated fatty acid (g) Polyunsaturated fatty acid (g) Cholesterol (mg) Carbohydrate (g) Sugar-total (g) Fibre (g) Retinol (pg) Carotene (pg) Retinol equivalent
Cases (n = 101) Mean (f S.D.)
Controls Mean
(n = 202) (i S.D.)
(952.8) (44.5) (39.7) (27.4)
3553.5 140.8 92.6 65.6
(1262.6) (53.3) (44.8) (36.1)
30.3
(14.0)
32.4
(15.6)
24.2
(11.3)
26.5
(13.4)
14.4* 480.2* 446.6*** 290.7** 27.8*** 2045.3 2432.3
(4.8) (202.5) (125.0) (91.1)
15.7 531.3 499.8 320.4 32.2 1939.7 2299.1
(6.6) (258.2) (169.3) (119.3) (13.1) (2110.6) (1865.6) (1698.2)
3201.7** 133.6 89.0 65.6
(8.1) (1468.0) (2104.1)
(P(P) Vitamin D (pg) 1-Tocopherol (mg) Thiamin (mg) Riboflavin (mg) Niacin (mg) Niacin equivalent
1740.1 (1270.1) 3.3 (2.0) 8.6*** (3.4) 1.6** (0.6) 2.8 (1.2) 18.2 (6.6)
1468.1 3.0 10.3 1.8 2.8 19.1
(1.9) (5.6) (0.8) (1.5) (8.5)
(mg) Vitamin B6 (mg) Folic acid ( pg) Vitamin B12 (pg) Vitamin C (mg) Sodium (mg) Potassium (mg) Calcium (mg) Phosphorus (mg) Magnesium (mg) Iron (mg) Zinc (mg)
28.2 (12.8) 1.9** (0.6) (61.2) 158.0 7.3* (4.7) 173.5 (91.9) 5506.1** (1631.1) 3843.1** (1233.2) 1128.7 (515.2) 1936.6** (623.3) (83.9) 219.0* 19.9*** (6.3) 8.7 (3.8)
28.9
(15.1)
2.2 167.0
(0.9) (97.5)
6.2 188.7 6026.7 4329.7 1221.3 2127.0 241.5
(6.0) (104.0) (2198.3) (1736.7) (546.7) (771.1) (109.9)
22.3 9.1
(8.3) (4.4)
None of the participants took vitamins or minerals *0.05
as supplements.
mean daily intake of z-tocopherol (P= 0.046) and iron (P’ 0.012). Considering that the ordinary scale is more suitable to logistic regression analysis than an interval one, and in order to examine the dose-response relationship, tertiles of average daily intake of nutrients were compared with low intake as the reference category. Odds ratios (OR) with 95% confidence interval (95% CI) for all nutrients are presented in Table 3. They are given as (a) unadjusted; (b) adjusted for energy; (c) adjusted for the nutrients which were significant in (a) or (b). The results of unadjusted univariate analysis showed that carbohydrate, sugar, retinol, retinol equivalent, rl-tocopherol, vitamin B12, sodium, potassium, calcium, phosphorus and iron were significantly related to prostate cancer only when their intake was high, with the exception of retino1 where both moderate and high intakes were indicated to be risk factors for prostate cancer. The overall trend was significant for all these variables (for phosphorous, the trend was of borderline significance).
Diet and Prostate Cancer Table 3. Odds ratios (9.5% conjidence
103
intervals) for prostate cancer,
by various nutrient intake levels
Unadjusted odds ratio Adjusted* odds ratio Adjusted** odds ratio Nutrients
Tertiles
Number of cases
Energy (kcal)
1 2 3
40 37 24
Protein
1 2 3
29 48 24
1 2 3
36 36 29
1 2 3
28 43 30
1 2 3
37 38 26
1 2 3
37 40 24
1 2 3
34 43 24
1 2 3
36 41 24
1 2 3
44 35 22
1 2 3
41 41 19
1 2 3
43 32 26
1 2 3
19 42 40
1 2 3
28 39 34
1 2 3
20 25 56
1 2 3
32 29 40
1 2 3
35 52 14
Fat--total
Fat-animal
Saturated fatty acids
Monounsaturated
fatty acids
Polyunsaturated fatty acids
Cholesterol
Carbohydrate
Sugar-total
Fibre
Retinol
Carotene
Retinol equivalent
Vitamin D
a-Tocopherol
(95% CI)
(95% CI)
1 0.87 (0.50-1.53) 0.77 (0.57-1.05) NS 1.58 (0.89-2.80) 0.90 (0.66-1.24) NS 0.99 (0.56-l 0.90 (0.67-l NS
.75) .ll)
1.51 (0.85-2.72) 1.01 (0.74-1.38) NS 0.97 (0.55-1.70) 0.84 (0.62-1.14) NS 1.02 (0.58-1.78) 0.82 (0.61-1.12) NS 1.25 (0.71-2.19) 1.19 (0.87-1.62) NS 1.07 (0.61-1.88) 0.82 (0.60-l. 11) NS 0.78 (0.45-1.37) 0.50 (0.27-0.92) P = 0.0280 0.94 (0.54-1.63) 0.46 (0.24-0.88) P = 0.0250 0.74 (0.42-1.31) 0.79 (0.59-l .06) NS 2.15 (1.13-4.06) 2.14 (1.12-4.07) P = 0.0285 1.31 (0.73-2.37) 1.01 (0.81-3.57) NS 1.18 (0.60-2.32) 2.80 (1.52-5.17) P = 0.0004 0.88 (0.48-1.61) 1.13 (0.84-1.50) NS 1 1.46 (0.85-2.53) 0.40 (0.20-0.81) P = 0.0284
(95% CI) 1 0.52 (0.15-1.81) 5.11 (1.37-19.01) NS
1 2.24 (1.00-5.04) 4.00 (1.42-11.31) NS 1 1.04 (0.47-2.23) 2.47 (1.18-5.19) NS 1 1.88 (0.96-3.67) 1.68 (0.97-2.91) NS 1 1.08 (0.55-2.14) 0.99 (0.52-1.88) NS 1 1.14 (0.58-2.26) 0.88 (0.50-1.58) NS 1 1.69 (0.83-3.44) 1.25 (0.69-2.26) NS 1 1.19 (0.65-2.16) 0.74 (0.49-1.14) NS 1 0.59 (0.27-1.30) 0.68 (0.33-1.40) NS 1 0.91 (0.50-1.64) 0.86 (0.52-1.43) NS 1 0.66 (0.32-1.37) 0.94 (0.52-1.69) NS 1 2.61 (1.33-5.10) 1.65 (1.16-2.36) P= 0.0046 1 1.49 (0.78-2.86) 1.38 (0.93-2.06) NS 1 1.52 (0.73-3.14) 2.09 (1.45-3.00) P = 0.0000 1 0.97 (0.52-1.84) 1.29 (0.91-1.81) NS 1 1.71 (0.90-3.25) 0.44 (0.24-0.79) NS
1.74 (0.65-5.64) 2.25 (2.47-59.51) NS 0.89 (0.31-2.54) 1.95 (0.68-5.57) NS 1 1.70 (0.74-3.90) 1.40 (0.62-3.15) NS 1 1.61 (0.63-4.03) 1.84 (1.03-11.18) NS 1 0.97 (0.33-2.51) 0.57 (0.16-2.10) NS 0.86 (0.33-2.23) 0.93 (0.37-2.32) NS 0.97 (0.46-2.06) 0.60 (0.33-1.10) NS 0.56 (0.20-1.60) 0.41 (0.90-6.61) NS 0.71 (0.32-1.58) 1.40 (0.57-3.46) NS 0.82 (0.27-2.50) 3.89 (1.36-11.13) P = 0.0584 1.22 (0.50-2.95) 0.69 (0.39-1.24) NS 1.87 (0.74-4.76) 1.24 (0.67-2.29) NS 1.06 (0.39-2.88) 1.67 (0.92-3.03) NS 0.81 (0.36-1.81) 0.70 (0.39-1.24) NS 1 1.17 (0.50-2.77) 0.22 (0.07-0.70) P = 0.0409 Continued
H.D.
104
Vlajinac
et al.
Table 3. Continued Thiamin
1
41 37 23
Riboflavin
1
30 42 29
Niacin
Niacin equivalent
1 2
34 63 4
1 3
36 33 32
2 3
35 40 26
2 3
26 38 37
2 3
20 24 57
n
‘
Vitamin B6
Folic acid
Vitamin B 12
Vitamin C 2
Sodium 2 3 Potassium
Calcium
Phosphorus
Magnesium
40 41 20
1 2
46 30 25
1 2 3
47 31 23
1 2 3
43 35 23
1 2 3
45 26 30
2 3
42 41 18
1 2 3
40 26 35
Iron
Zinc
36 33 32
1 0.88 (0.50-1.53) 0.75 (0.56-1.03) NS
1 1.21 (0.60-2.47) 1.37 (0.72-2.60) NS
1 1.65 (0.4885.60) 1.10 (0.34-3.57) NS
1.32 (0.74-2.35) 1.01 (0.74-1.37) NS
1.82 (0.88-3.78) 1.44 (0.88-2.36) NS
1.61 (0.49-5.29) 0.49 (0.15-1.64) NS
1.00 (0.60-1.68) 0.74 (0.41-1.35) NS
1.51 (0.76-3.00) 1.05 (0.34-3.19) NS
0.88 (0.49-1.57) 0.96 (0.71-1.28) NS
1.11 (0.55-2.26) 1.22 (0.74-2.01) NS
1.05 (0.42-2.61) 0.21 (0.02-2.62) NS 1 0.86 (0.28-2.59) 1.21 (0.37-3.93) NS
1.09 (0.62-1.92) 0.86 (0.63-1.16) NS
1.27 (0.66-2.42) 1.01 (0.59-1.75) NS
1.38 (0.76-2.51) 1.19 (0.88-1.62) NS
1.75 (0.89-3.43) 1.83 (1.19-2.83) P= 0.0158
1.18 2.85 P 1 0.88 0.94
(0.60-2.34) (1.55-5.25) = 0.0003
1.54 (0.73-3.22) 2.35 (1.61-3.45) P = 0.0000
(0.49-1.57) (0.70-1.25) NS
0.54 (0.26-1.14) 0.90 (0.48-1.69) NS
0.97 (0.56-1.68) 0.50 (0.26-0.94) P= 0.0415
0.92 (0.50-1.70) 1.16 (0.79-1.70) NS 1 0.76 (0.36-1.62) 1.22 (0.54-2.78) NS
0.61 (0.35-1.09) 0.54 (0.30-0.98) P = 0.0373
0.55 (0.26-1.13) 0.68 (0.38-1.20) NS
0.53 (0.16-1.69) 0.92 (0.33-2.60) NS
0.62 (0.35-1.09) 0.49 (0.27-0.90) P= 0.0171
0.59 (0.31-1.13) 0.59 (0.34-1.02) NS
0.51 (0.19-1.34) 0.37 (0.14-0.94) P= 0.0346
0.79 (0.45-l .38) 0.54 (0.30-1.00) P = 0.0507
0.68 (0.32-1.44) 0.86 (0.41-1.80) NS
0.99 (0.28-3.51) 1.69 (0.44-6.43) NS
0.56 (0.31-1.01) 0.80 (0.60-1.07) NS
0.59 (0.30-1.14) 0.83 (0.53-1.32) NS 1 0.80 (0.39-1.62) 0.47 (0.22-1.00) P = 0.0486 1 0.75 (0.38-1.48) 1.31 (0.81-2.13) NS
0.54 (0.18-1.58) 0.36 (0.13-0.99) NS
0.92 (0.53-1.59) 0.43 (0.22-0.82) P= 0.0145 0.64 (0.35-1.16) 0.96 (0.70-1.24) NS
*Adjusted for energy. **Adjusted for nutrients which are significant when unadjusted or adjusted for energy. f value, for linear trend of odds ratio. NS, non-significant.
0.98 (0.38-2.54) 0.78 (0.23-2.69) NS 1 1.32 (0.55-3.15) 1.38 (0.62-3.08) NS 1 1.57 (0.62-3.97) 2.02 (1.06-3.85) P = 0.0229
0.54 (0.22-1.31) 1.56 (0.64-3.84) NS
0.47 (0.17-1.25) 0.44 (0.17-1.18) P= 0.0215 0.41 (0.14-1.16) 0.81 (0.28-2.34) NS
Diet and Prostate Table 4.
Variables
sign$cantly
related to prostate
cancer
after
adjustment for possible confounders*
Tertiles
Variable Protein
Saturated
fatty acid
Fibre
a-Tocopherol
Vitamin
B 12
Calcium
Iron
1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
*Energy, protein, fat-total, sugar-total, fibre, retinol, acid, vitamin B12, sodium, nesium and iron.
Odds ratio 1 1.71 13.54 1 1.54 3.63 1 0.83 4.02 1 1.10 0.15 1 1.74 2.07 1 0.35 0.37 1 0.40 0.34
95% confidence interval
Significance of linear trend
0.56-5.24 2.38-77.13
NS
0.60-3.95 1.03-12.79
NS
0.27-2.51 1.38-11.73
0.0564
0.45-2.74 0.05-0.53
0.0117
0.66-4.56 1.08-3.97
0.0200
0.12-1.00 0.14-0.99
0.0189
0.14-1.11 0.12-0.95
0.0089
saturated fatty acids, carbohydrate, retinol equivalent, a-tocopherol, folic potassium, calcium, phosphorus, mag-
After adjustment for energy, a significant relationship with prostate cancer was found with high protein intake, fat-total, retinol (moderate and high intake), retinol equivalent, a-tocopherol, folic acid, vitamin B12 and iron. A linear trend was significant for retinol, retinol equivalent, folic acid, vitamin B12 and iron. After adjustment for nutrient variables significantly related to prostate cancer using unadjusted or adjusted for energy, a significant relationship with prostate cancer was found for high energy intake, protein, saturated fatty acids, fiber, LXtocopherol, vitamin B12, calcium and magnesium. A linear trend was significant for fibre, a-tocopherol, vitamin B12 and calcium. In Table 4 are presented odds ratios and 95% confidence intervals for nutrients sigmficantly related to prostate cancer after adjustment for all factors significantly associated with prostate cancer according to previous analyses. High protein intake, saturated fatty acid, fibre and vitamin B12 appeared to be risk factors. A protective effect was found for a high intake of a-tocopherol, calcium and iron. When all these variables, significantly related to prostate cancer according to previous analyses, were included in the
Table 5. Risk factors for prostate cancer-multivariate
Nutrient Retinol equivalent (mg) Vitamin B 12 (mg) Iron (mg) Constant
B 0.4940 0.6258 -0.9279 -1.3905
Cancer
105
model of logistic regression step by step analysis, risk factors for prostate cancer appeared to be retinol equivalent and vitamin B12, and a protective factor was iron (Table 5). The results did not change when other non-dietary factors significantly related to prostate cancer were included in the model (occupational physical activity during the year preceding the disease; specific occupational exposure, that is, exposure to asbestos, steel, dyes and lacquers, bitumen, pitch, iron, nickel, lead, fertiliser and other potentially harmful agents; nephrolithiasis; “other” diseases such as chronic bronchitis, chronic rheumatic diseases, hypertension, cardiomyopathia, diabetes mellitus, renal diseases, eye diseases and tuberculosis; greater number (2 3) of brothers; greater number (2 8) of sexual partners [6]. The reason why retinol equivalent was not found to be related to prostate cancer when adjustment was made for vitamin B12 was the high colinearity between these two variables (Pearson coefficient of linear correlation r = 0.7636, P= 0.0000). Since some investigations suggested that the vitamin A effect depends on fat intake, we observed the relationship of prostate cancer with retinol equivalent and carotene in participants whose average fat intake was low (lower than the median value of the control group) and those whose average fat consumption was high (> median values for control group). Total fat and animal fat intake were observed separately. Retinol equivalent was directly related to prostate cancer irrespective of the quantity and type of fat consumed, although subjects whose animal fat intake was > median value had much higher relative risk in comparison with those whose animal fat intake was below median (OR= 4.68, 95% CI = 3.80-5.57 and OR = 1.90, 95% CI = 1.20-2.61, respectively). In those whose total fat intake was > median, the odds ratio for retinol equivalent was 3.04 (2.08-4.00) and in participants whose total fat intake was below the median, the odds ratio was 3.53 (2.86-4.19). For carotene intake such an association was not present. DISCUSSION The main dietary component which has been associated with prostate cancer risk is fat [l]. Ecological studies and the majority of case-control studies [7-91 have shown a strong positive association. However, cohort studies [ 1 O-l 21 have not been consistent in their findings with regard to dietary fat. Snowdon and associates [lo] found that prostate cancer mortality was positively associated with the consumption of milk, cheese, eggs and meat. In the study of Giovannucci and associates [l 11, total fat was directly related to risk of advanced prostate cancer-the association being due primarily to animal fat. In contrast, in Severson and associates’ study [12], there was no indication that heavy consumption of fat increased the risk of prostate can-
analysis (cases were divided at each tertile level of controls)
Standard 0.2484 0.2502 0.1993 0.4282
error
P value
Odds ratio
95% confidence interval
0.0467 0.0124 0.0000 0.0012
1.64 1.87 0.40
1 .Ol-2.67 1.15-3.05 0.27-0.58
106
H.D. Vlaiinac
cer, and Hsing and associates [13] also observed no association with consumption of meat, eggs or dairy products. It has been hypothesised that dietary fat affects cancer occurrence via alteration in the hormone environment. Some experimental data [ 14, 151 and analysis of hormone levels in the human population provide some corroborative information. A Western diet fed to black South African men increased the urinary excretion of oestrogen and androgens, while the excretion of oestrogen and androgens decreased in black North American men fed a vegetarian diet [16]. Bishop and associates [17] correlated differences in sex hormone levels in 171 twin pairs with differences in reported dietary habits. They found differences in fat intake correlated with differences in testosterone levels. Which fat component in particular may be responsible for the association with prostate cancer has not yet been established [l]. The association with saturated fat observed in our study had been most frequently found [ 18, 191 or suggested by the association with animal fat and with meat or dairy products [20]. However, West and associates [21] found a relationship with saturated and unsaturated fat. Giovannucci and associates [ 1 l] found that saturated fat, unsaturated fat and a-linolenic acid were associated with advanced prostate cancer risk, but only an association with cc-linolenic acid persisted when saturated fat, unsaturated fat, linolenic acid and a-linolenic acid were modelled simultaneously. Data about the relationship of vitamin A or its precursors, especially beta-carotene, to prostate cancer risk, are also inconsistent both in epidemiological and in animal studies [I]. Our findings that retinol and retinol equivalent are risk factors for prostate cancer, with an odds ratio of around 2.00, is in accordance with many case-control [22-251 studies and some cohort studies [ 12, 13, 261. Studies in which an inverse association has been found are also numerous [20, 27, 281. Methodological differences in study design and dietary survey methods might, in part, explain these inconsistencies, but opposite results were obtained even in studies using the same method for dietary data collection. The suggestion that the effect of vitamin A or its precursors depends on fat intake [28] is not supported by our data, showing a positive effect of retinol equivalent and no effect of carotene, independently of quantity and type of fat consumed. Nevertheless, it is possible that moderate amounts of vitamin A might protect against prostate cancer, while large amounts could enhance the risk [28]. According to a recent report, a trial in the United States involving 18 000 smokers has been discontinued 2 years early after initial results showed that taking vitamin A and beta-carotene as supplements may increase the risk of cancer [29]. Data on the effect of energy intake are also inconsistent. Severson and associates [ 121 did not find that total energy intake caused an increased risk of prostate cancer. According to the study of Rose and associates [30], prostate cancer mortality rates were only weakly associated with total calorie intake in the 28 countries, due to a strong positive correlation with calories of animal origin. In the study of West and associates [21], energy intake had an odds ratio of 2.5 (95% CI= 1.0-6.5). In our study the highest tertile of total energy intake was a risk factor for prostate cancer after adjustment for possible confounders including protein, tota fat, carbohydrate and total sugar.
et al.
The positive association between prostate cancer and protein intake found in our investigation was noted in studies of Kolonel and associates [31], Heshmat and associates [32] and Metlin and associates [20], but was not found in other studies [12, 271. The effect of total caloric intake as well as protein would most likely be through hormones. The positive relationship between prostate cancer and fibre intake found in this investigation would also most likely be through hormones. Lubin and associates [33] found that breast cancer risk decreases with a higher consumption of food rich in fibre, even in those with a high intake of fat and protein. Adlercreutz [34] suggested that breast cancer risk may be favourably influenced by lignans, precursors of which occur in the diet and are associated with fibre component. He hypothesised that the fibre had an anti-oestrogen effect. Since oestrogen has a beneficial effect in the treatment of prostate cancer, it is postulated that a decrease in the oestrogen level could increase the risk of prostate cancer [35]. At the same time, the lower risk of prostate cancer in cirrhotic patients could be the result of hyperoestrogenism that is present in subjects with cirrhosis [36, 371. There is no literature data on vitamin B12 as a risk factor for prostate cancer. Vitamin B12 is involved in all metabolic processes, but at present we have no adequate explanation for its positive connection with prostate cancer. The protective effect of vitamin E and iron, observed in could have a plausible explanation. the present study, Vitamin E is known as an intracellular antioxidant [38]. It also stimulates T cells and increases immunological reactions [39]. In vitro and in viva vitamin E inhibits the nitrosamine synthesis of N-nitroso compounds, and nitrosamide, which are known to be carcinogenic in laboratory animals [40]. The protective effect of vitamin E has been reported for different cancers including colon cancer, head and neck cancers, cervical cancer [41] and breast cancer [42]. A study is ongoing to investigate whether S(-tocopherol will increase the DNA repair capacity in patients at risk of head, neck and colon cancer [43]. The fact that iron is involved in oxidation-reduction processes makes it plausible that iron may decrease the risk of cancer by inhibition of free radical reactions. There are no data about the relationship of prostate cancer and calcium intake, but the protective effect of calcium has been reported for some other cancers [44, 451. Two clinical trials are ongoing to investigate the effect of calcium supplementation in colon cancer chemoprevention [46] and in preventing recurrences of neoplastic polyps of the large bowel [47]. The relationship of prostate cancer with vitamin B12, vitamin E, fibre, calcium and iron should be corroborated by other investigations.
Nomura AMY, Kolonel LN. Prostate cancer: a current perspective. Am 3 Epidemiol 1991, 13, 200-227. Jain D. Diet history: questionnaire and interview techniques used in some retrospective studies of cancer. 3 Am Diet Assoc 1989,89, 1647-1652. Kaic-Rak A, Antonic K. Tables of Food Items Composition. Zagreb, Public Health of Croatia, 1990. Breslow NE, Day NE. Statistical Methods in Cancer Research. Volume I-The Analysis of Case-Control Studies. Lyon, IARC, 1980, 248-279.
Diet
and
Prostate
5. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York, Wiley-Interscience 1989, 82-134; 187-215. 6. IliC M, Vlajinac H, Marinkovic J. Case-control study of risk factors for prostate cancer. Br J Cancer 1996, 74, 1682-1685. 7. Berg JW. Can nutrition explain the pattern of international epidemiology of hormonal-dependent cancers? Cancer Res 1975,35,3345-3350. 8. Kolonel LN, Hankin JH, Lee J, Chu SY, Nomura AMY, Hinds MW. Nutrient intakes in relation to cancer incidence in Hawaii. Br3 Cancer 198 1, 44, 332-339. 9. Slattery ML, Schumacher MC, West DW, Robison LM, French TK. Food-consumption trends between adolescent and adult years and subsequent risk of prostate cancer. Am 3 Clin Nua 1990, 52, 752-757. 10. Snowdon DA, Phillips RL, Choi W. Diet, obesity, and risk of fatal prostate cancer. Am 3 Epidemiol 1984, 120, 244-250. 11. Giovannucci E, Rimm :EB, Colditz GA, et al. A prospective study of dietary fat and risk of prostate cancer. 3 Nat1 Cancer znst 1993,85, 1571-1570. 12. Severson RK, Nomura AMY, Grove JS, Stemmennann GN. A prospective study of demographics, diet, and prostate cancer among men of Japanese ancestry in Hawaii. Cancer Res 1989,49, 1857-1860. 13. Hsing AW, McLaughlin JK, Schuman LM, et al. Diet, tobacco use and fatal prostate cancer: results from the Lutheran brotherhood cohort study. Cancer Res 1990, 50, 6836-6840. 14. Pollard M, Luckert PH. Promotional effects of testosterone and dietary fat on prostate carcinogenesis in genetically susceptible rat. Prostate 1985, 6, l-5. 15. Carroll KK, Noble RL. Dietary fat in relation to hormonal induction of mammary and prostatic carcinoma in Nb rats. Carcinogenesis 1987, 8, 85-853. 16. Hill P, Garbaczewski L, Helman I’, Walker ARE, Games H, Wynder EL. Environmental factors and breast and prostatic cancer. Cancer Res 1981, 41, 3817-3818. 17. Bishop DT, Meikie AW, Slattery ML, Stringham JD, Ford MH, West DW. The effect of nutritional factors on sex hormone levels in male twins. Gen Epidemiol 1988, 5, 43-59. 18. Kolonel LN, Yoshizawa CN, Hankin JH. Diet and prostatic cancer: a case-control study in Hawaii. Am 3 Epidemiol 1988, 127, 999-1012. 19. Hankin JH, Zhao Ll’, Wilkens LR, Kolonel LN. Attributable risk of breast, prostate and lung cancer in Hawaii due to saturated fat. Cancer Causes Control 1992, 3, 17-23. 20. Mettlin C, Selenskas S, Natarajan N, Huben R. Beta-carotene and animal fats and their relationship to prostate cancer risk. A case-control study. Cancer 1989, 64, 605-612. 21. West DW, Slattery ML, Robinson LM, French TK, Mahoney AW. Adult dietary intake and prostate cancer risk in Utah: a case-control study with special emphasis on aggressive tumors. Cancer Causes Control 19511, 2, 85-94. 22. Graham S, Haughey B, Marshall J, et al. Diet in the epidemiology of carcinoma of the prostate gland. 3 NatZ Cancer Inst 1983, 70, 687-692. 23. Heshmat MY, Kani L, Kovi J, et al. Nutrition and prostate cancer: a case-control study. Prostate 1985, 6, 7-17. 24. Kolonel LN, Hankin JH, Yoshizawa CN. Vitamin A and prostate cancer in elderly man: enhancement of risk. Cancer Res 1987,47, 2982-2985. 25. Talamini R, La Vecchia C, Decarli A, Negri E, Franceschi S. Nutrition, social factors, and prostatic cancer in a Northern Italian population. Br 3 Cancer 1986, 55, 8 17-82 1. 26. Paganini-Hill A, Chao A, Ross RK, Henderson BE. Vitamin A, beta carotene and the risk of cancer: a prospective study. 3 NatZ Cancer Inst 1987, 79, 443-448. 27. Ross RK, Shimicu H, Paganini-Hill A. Case-control studies of prostate cancer in blacks and whites in southern California. 3 NatZ Cancer Inst 1987, 78, 869-874.
28.
29. 30.
3 1.
32. 33.
34.
35. 36. 37.
38. 39.
40.
41.
42.
43.
44. 45.
46.
47.
Cancer
107
Ohno Y, Yoshida 0, Oishi K, Okada K, Yamabe H, Schoeder FH. Dietary beta carotene and cancer of the prostate: a casecontrol study in Kyoto, Japan. Cancer Res 1988, 48, 13311336. Cookson C. Cancer-risk fears halt US tests on vitamin pills for smokers. Financial Times 1996, January 20121, 24. Rose DP, Boyar AI’, Wynder EL. International comparisons of mortality rates for cancer of the breast, ovary, prostate and colon, and per capita food consumption. Cancer 1986, 58, 2363-237 1. Kolonel LN, Hankin JH, Lee J, Chu SY, Nomura AMY, Ward Hinds M. Nutrient intakes in relation to cancer incidence in Hawaii. Br3 Cancer 1981, 44, 332-339. Heshmat MY, Kani L, Kovi J, et al. Nutrition and prostate cancer: a case-control study. Prostate 1985, 6, 7-17. Lubin F, Wax Y, Modan B. Role of fat, animal protein and dietary fiber in breast cancer etiology. A case-control study. 3 Nat1 Cancer Inst 1986, 77, 605-612. Adlercreutz H. Does fiber-rich food containing animal lignan precursors protect against both colon and breast cancer? An extension of the “fiber hypothesis”. Gasroenterology 1984, 86, 761-766. Resnik M. Hormonal therapy in prostatic carcinoma. UroZogy 1984, 24, 18-20. Glantz MG. Cirrhosis and carcinoma of the prostate gland. 3 Ural 1964,91, 291-293. Chopra IJ, Tulchinsky D, Greenway FL. Estrogen-androgen imbalance in hepatic cirrhosis. Ann Intern Med 1973, 79, 19% 203. Bieri JG, Corash L, Hubbard VS. Medical use of vitamin E. N EngZJ Med 1983, 308, 1063-1071. Raletic-Dragicevic J, Djulizibaric J. Can vitamins have preventive effect on cancer? Information About Cuncerogenics (in Serbian) 1985, 2, 10-12. Choi BCK. N-nitroso compounds and human cancer: a molecular epidemiologic approach. Am 3 Epidemiol 1985, 121, 737-743. Verreault R, Chu J, Mandelson M, Shy K. A case-control study of diet and invasive cervical cancer. IntJ Cancer 1989,43, 1050-1054. Schrauzer GN, White DA, Schneider SJ. Cancer mortality correlation studies. III Statistical association with dietary selenium intakes. Bioinorg Chem 1977, 7, 23-34. Berwick M, Schantz SP, Liebes L, Shike M. Vitamin E and DNA repair in patients at risk for head and neck cancer and colon cancer. In Sankaranaryanan R, Wahrendorf J, Demaret E in collaboration with Becker N, eds. Directoy of On-Going Research in Cancer Epidemiology. Lyon, IARC, 1996, 463. Rosen M, Nystrom L, Wall S. Diet and cancer mortality in the counties of Sweden. Am 3 Epidemiol 1988, 127, 42-49. Meyer F, White E. Alcohol and nutrients in relation to colon cancer in middle-aged adults. Am 3 Epidemiol 1993, 138, 225226. Wargovich MJ, Steinbach G, Winn R, Lee J, Levin B. Calcium and aspirin prevention trial. In Sankaranaryanan R, Wahrendorf J, Demaret E in collaboration with Becker N, eds. Directoy of On-Going Research in Cancer Epidemiology. Lyon, IARC, 1996, 437. Baron JA, Rothstein R, Beck G, er al. Calcium in the prevention of neoplastic polyps. In Sankaranaryanan R, Wahrendorf J, Demaret E in collaboration with Becker N, eds. Directoy of On-Going Research in Cancer Epidemiology. Lyon, IARC, 1996, 429.
Acknowledgement-Funded Technology of Serbia through
by the Ministry contract No. 8774,
for Science 1991-5.
and
Cancervol. 33, No. 1, pp. 101-107, 1997 c: 1997 Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0959.8049197 $17.00+0.00
Pergamon PII: SO959-8049(96)00373-S
Original Paper Diet and Prostate Cancer: a Case-Control H.D. Vlajinac,’ ‘Institute
of Epidemiology,
‘Institute
of Social Belgrade;
J.M. MarinkoviC,2
School
Medicine, and School
of Medicine,
Statistics
Belgrade
and Health
of Medicine,
M.D.
University,
Research,
Kragujevac
Ilit
School
University,
Study
and N.I. Kocev’ Visegradska
26,
11000,
of Medicine,
Belgrade
Kragujevac,
Yugoslavia
Belgrade; University,
A case-control study., performed in two towns of Serbia (Yugoslavia) from 1990 to 1994, comprised 101 patients with histologically confirmed prostate cancer and 202 hospital controls individually matched by age (k2 years), hospital admittance and place of residence. Dietary information was obtained by using a standard questionnaire. After adjustment for possible confounders, risk factors for prostate cancer appeared to be the highest tertile of protein (odds ratio (OR) = 13.54, 95% confidence interval (CI) = 2.38-77.13), saturated fatty acid (OR = 3.63, 95% CI = 1.03-12.79), fibre (OR = 4.02, 95% CI = 1.38-11.73), and vitamin B12 intake (OR = 2.07, 95% CI = 1.08-3.97); a protective effect was found for the highest tertile of a-tocopherol (OR = 0.15, 95% CI = 0.05-0.53), calcium (OR = 0.37, 95% CI = 0.14-0.99) and iron intake (OR = 0.34, 95% CI = 0.12-0.95). There were significant (PC 0.05) linear trends in the odds ratios for c+tocopherol, vitamin B12, calcium and iron. According to logistic regression step by step analysis, risk factors for prostate cancer were dietary intake of retinol equivalent (OR = 1.64, 95% CI = 1.01-2.67) and vitamin B12 (OR = 1.87, 95% CI = 1.15-3.05), and a protective effect was found for dietary intake of iron (OR = 0.40, 95% CI = 0.27-0.58). 0 19!#7 Published by Elsevier Science Ltd. All rights reserved. Key words: prostate cfancer, diet, case-control EurJ
Cancer,
Vol. 33, No.
1, pp. 101-107,
study
1997
INTRODUCTION LrrrLEISknown about the aetiology of prostate cancer. A
1994. Of 141 patients with histologically confirmed clinical prostate cancer, 12 patients could not be interviewed as they gave incorrect addresses, 9 patients refused to participate, 10 patients were too ill to be interviewed and 9 patients died before we were able to contact them. The final group consisted of 101 prostate cancer patients. For each case, two hospital controls (202 controls in total) were chosen among patients confirmed as having neither prostate cancer nor other prostate diseases. Those with other malignancies were also excluded. The majority of controls-154 patients, were treated at the hospital because of physical injuries, 11 had asthma, 8 had pneumonia, 8 had a peptic ulcer, 7 had cholecistitis, 6 had angina pectoris, 4 had cirrhosis, 3 had pleuritis and 1 had pancreatitis. All selected controls were interviewed-no one refused to participate. Cases and controls were individually matched by age (f2 years), hospital admittance and place of residence. Demographic, epidemiological and dietary data were obtained using a standard questionnaire. Dietary information was obtained by a quantitative history approach in which subjects were asked about their usual frequency of
number of risk factors h.ave been identified in past epidemiological studies, although the findings have not always been reproducible. The plausibility of diet having a major role in the aetiology of ptostate cancer is supported by certain observations based on descriptive epidemiology of the disease [l], but the results of analytical epidemiological studies have been inconsistent. This report describes results relating to dietary factors, obtained in a case-control study carried out in Serbia. A quantitative dietary history was used to estimate the weekly intake of different nutrients. Results related to non-dietary factors will be reported separately.
MATERIALS AND METHODS The study was conducted in two towns of Central Serbia (Kragujevac and Cuprija) from January 1990 to December Correspondence to H. Vlajinac. Received 20 Feb. 1996; revised 26 Jul. 1996; accepted 3 Sep. 1996. 101
102
H.D. Vlajinac
intake and portion size of a list of 150 food items including alcoholic beverages. The technique was similar to the one used by Jain and associates [2], although somewhat modified and adapted to suit the Serbian diet. Participants were asked how many months per year they used each food item and how often (daily/weekly/monthly/yearly). To facilitate quantification of intakes, 21 photographs of food items in three different portion sizes of known quantity were used in the interview. As measurements of consumption, standard cups, spoons, slices etc. were used. The referent period was 10 years before the disease. If significant changes in the diet had happened during the period observed, dietary habits before the change were recorded. Subjects were also questioned about their intake of vitamin and mineral supplements. From data obtained, the total daily amount of consumption for each food item was calculated and then finally converted into nutrients using food consumption tables [3]. For statistical analysis of data, univariate and multivariate logistic regression methods were used [4, 51. Tertiles of average daily intake of control group were used as a basis for comparison. A test for linear trend in risk was calculated as proposed by Breslow and Day [4]. RESULTS Table 1 describes basic demographic and personal characteristics of the case and control subjects. There were no notable differences between the groups. Table 2 gives the mean values of daily intake of nutrients. According to univariate logistic regression analysis, the mean intake of carbohydrates, sugar, fibre, tocopherol, thiamin, vitamin B6, sodium, potassium, phosphorus and iron and the mean total energy were significantly different between cases and controls. These variables, together with the mean inake of polyunsaturated fatty acids, cholesterol, vitamin B12 and magnesium, related to prostate cancer at a significant level of P < 0.10 entered into the model of multivariate analysis. According to multivariate logistic regression analysis, independent positive association with prostate cancer was found for the mean daily intake of vitamin B12 (P= 0.000) and a negative association was found for the
Table
1. Characteristics
of prostate
cancer
patients
and
their
controls
Characteristic Mean age (years) Ever married (%) Education (years) % o-4 5-12 >12 Occupation (%) Farmers Workers Clerks Place of residence (%) Urban Rural Mean height (cm) Mean weight (kg) Mean body mass index
Cases (n = 101)
Controls (I? = 202)
70.5
71.5
100.0
99.5
58.4 37.6 4.0
64.8 29.2 5.9
37.6 44.6 17.6
35.6 50.0 14.4
48 52 172.9 75.3 25.2
48 52 174.2 74.9 24.3
et al.
Table 2. Mean
values
(standard
deviation)
of daily intake
of
dietary nutrients
Nutrient
(unit)
Energy (kcal) Protein (g) Fat-total (g) Fat-animal (g) Saturated fatty acid (g) Monounsaturated fatty acid (g) Polyunsaturated fatty acid (g) Cholesterol (mg) Carbohydrate (g) Sugar-total (g) Fibre (g) Retinol (pg) Carotene (pg) Retinol equivalent
Cases (n = 101) Mean (f S.D.)
Controls Mean
(n = 202) (i S.D.)
(952.8) (44.5) (39.7) (27.4)
3553.5 140.8 92.6 65.6
(1262.6) (53.3) (44.8) (36.1)
30.3
(14.0)
32.4
(15.6)
24.2
(11.3)
26.5
(13.4)
14.4* 480.2* 446.6*** 290.7** 27.8*** 2045.3 2432.3
(4.8) (202.5) (125.0) (91.1)
15.7 531.3 499.8 320.4 32.2 1939.7 2299.1
(6.6) (258.2) (169.3) (119.3) (13.1) (2110.6) (1865.6) (1698.2)
3201.7** 133.6 89.0 65.6
(8.1) (1468.0) (2104.1)
(P(P) Vitamin D (pg) 1-Tocopherol (mg) Thiamin (mg) Riboflavin (mg) Niacin (mg) Niacin equivalent
1740.1 (1270.1) 3.3 (2.0) 8.6*** (3.4) 1.6** (0.6) 2.8 (1.2) 18.2 (6.6)
1468.1 3.0 10.3 1.8 2.8 19.1
(1.9) (5.6) (0.8) (1.5) (8.5)
(mg) Vitamin B6 (mg) Folic acid ( pg) Vitamin B12 (pg) Vitamin C (mg) Sodium (mg) Potassium (mg) Calcium (mg) Phosphorus (mg) Magnesium (mg) Iron (mg) Zinc (mg)
28.2 (12.8) 1.9** (0.6) (61.2) 158.0 7.3* (4.7) 173.5 (91.9) 5506.1** (1631.1) 3843.1** (1233.2) 1128.7 (515.2) 1936.6** (623.3) (83.9) 219.0* 19.9*** (6.3) 8.7 (3.8)
28.9
(15.1)
2.2 167.0
(0.9) (97.5)
6.2 188.7 6026.7 4329.7 1221.3 2127.0 241.5
(6.0) (104.0) (2198.3) (1736.7) (546.7) (771.1) (109.9)
22.3 9.1
(8.3) (4.4)
None of the participants took vitamins or minerals *0.05
as supplements.
mean daily intake of z-tocopherol (P= 0.046) and iron (P’ 0.012). Considering that the ordinary scale is more suitable to logistic regression analysis than an interval one, and in order to examine the dose-response relationship, tertiles of average daily intake of nutrients were compared with low intake as the reference category. Odds ratios (OR) with 95% confidence interval (95% CI) for all nutrients are presented in Table 3. They are given as (a) unadjusted; (b) adjusted for energy; (c) adjusted for the nutrients which were significant in (a) or (b). The results of unadjusted univariate analysis showed that carbohydrate, sugar, retinol, retinol equivalent, rl-tocopherol, vitamin B12, sodium, potassium, calcium, phosphorus and iron were significantly related to prostate cancer only when their intake was high, with the exception of retino1 where both moderate and high intakes were indicated to be risk factors for prostate cancer. The overall trend was significant for all these variables (for phosphorous, the trend was of borderline significance).
Diet and Prostate Cancer Table 3. Odds ratios (9.5% conjidence
103
intervals) for prostate cancer,
by various nutrient intake levels
Unadjusted odds ratio Adjusted* odds ratio Adjusted** odds ratio Nutrients
Tertiles
Number of cases
Energy (kcal)
1 2 3
40 37 24
Protein
1 2 3
29 48 24
1 2 3
36 36 29
1 2 3
28 43 30
1 2 3
37 38 26
1 2 3
37 40 24
1 2 3
34 43 24
1 2 3
36 41 24
1 2 3
44 35 22
1 2 3
41 41 19
1 2 3
43 32 26
1 2 3
19 42 40
1 2 3
28 39 34
1 2 3
20 25 56
1 2 3
32 29 40
1 2 3
35 52 14
Fat--total
Fat-animal
Saturated fatty acids
Monounsaturated
fatty acids
Polyunsaturated fatty acids
Cholesterol
Carbohydrate
Sugar-total
Fibre
Retinol
Carotene
Retinol equivalent
Vitamin D
a-Tocopherol
(95% CI)
(95% CI)
1 0.87 (0.50-1.53) 0.77 (0.57-1.05) NS 1.58 (0.89-2.80) 0.90 (0.66-1.24) NS 0.99 (0.56-l 0.90 (0.67-l NS
.75) .ll)
1.51 (0.85-2.72) 1.01 (0.74-1.38) NS 0.97 (0.55-1.70) 0.84 (0.62-1.14) NS 1.02 (0.58-1.78) 0.82 (0.61-1.12) NS 1.25 (0.71-2.19) 1.19 (0.87-1.62) NS 1.07 (0.61-1.88) 0.82 (0.60-l. 11) NS 0.78 (0.45-1.37) 0.50 (0.27-0.92) P = 0.0280 0.94 (0.54-1.63) 0.46 (0.24-0.88) P = 0.0250 0.74 (0.42-1.31) 0.79 (0.59-l .06) NS 2.15 (1.13-4.06) 2.14 (1.12-4.07) P = 0.0285 1.31 (0.73-2.37) 1.01 (0.81-3.57) NS 1.18 (0.60-2.32) 2.80 (1.52-5.17) P = 0.0004 0.88 (0.48-1.61) 1.13 (0.84-1.50) NS 1 1.46 (0.85-2.53) 0.40 (0.20-0.81) P = 0.0284
(95% CI) 1 0.52 (0.15-1.81) 5.11 (1.37-19.01) NS
1 2.24 (1.00-5.04) 4.00 (1.42-11.31) NS 1 1.04 (0.47-2.23) 2.47 (1.18-5.19) NS 1 1.88 (0.96-3.67) 1.68 (0.97-2.91) NS 1 1.08 (0.55-2.14) 0.99 (0.52-1.88) NS 1 1.14 (0.58-2.26) 0.88 (0.50-1.58) NS 1 1.69 (0.83-3.44) 1.25 (0.69-2.26) NS 1 1.19 (0.65-2.16) 0.74 (0.49-1.14) NS 1 0.59 (0.27-1.30) 0.68 (0.33-1.40) NS 1 0.91 (0.50-1.64) 0.86 (0.52-1.43) NS 1 0.66 (0.32-1.37) 0.94 (0.52-1.69) NS 1 2.61 (1.33-5.10) 1.65 (1.16-2.36) P= 0.0046 1 1.49 (0.78-2.86) 1.38 (0.93-2.06) NS 1 1.52 (0.73-3.14) 2.09 (1.45-3.00) P = 0.0000 1 0.97 (0.52-1.84) 1.29 (0.91-1.81) NS 1 1.71 (0.90-3.25) 0.44 (0.24-0.79) NS
1.74 (0.65-5.64) 2.25 (2.47-59.51) NS 0.89 (0.31-2.54) 1.95 (0.68-5.57) NS 1 1.70 (0.74-3.90) 1.40 (0.62-3.15) NS 1 1.61 (0.63-4.03) 1.84 (1.03-11.18) NS 1 0.97 (0.33-2.51) 0.57 (0.16-2.10) NS 0.86 (0.33-2.23) 0.93 (0.37-2.32) NS 0.97 (0.46-2.06) 0.60 (0.33-1.10) NS 0.56 (0.20-1.60) 0.41 (0.90-6.61) NS 0.71 (0.32-1.58) 1.40 (0.57-3.46) NS 0.82 (0.27-2.50) 3.89 (1.36-11.13) P = 0.0584 1.22 (0.50-2.95) 0.69 (0.39-1.24) NS 1.87 (0.74-4.76) 1.24 (0.67-2.29) NS 1.06 (0.39-2.88) 1.67 (0.92-3.03) NS 0.81 (0.36-1.81) 0.70 (0.39-1.24) NS 1 1.17 (0.50-2.77) 0.22 (0.07-0.70) P = 0.0409 Continued
H.D.
104
Vlajinac
et al.
Table 3. Continued Thiamin
1
41 37 23
Riboflavin
1
30 42 29
Niacin
Niacin equivalent
1 2
34 63 4
1 3
36 33 32
2 3
35 40 26
2 3
26 38 37
2 3
20 24 57
n
‘
Vitamin B6
Folic acid
Vitamin B 12
Vitamin C 2
Sodium 2 3 Potassium
Calcium
Phosphorus
Magnesium
40 41 20
1 2
46 30 25
1 2 3
47 31 23
1 2 3
43 35 23
1 2 3
45 26 30
2 3
42 41 18
1 2 3
40 26 35
Iron
Zinc
36 33 32
1 0.88 (0.50-1.53) 0.75 (0.56-1.03) NS
1 1.21 (0.60-2.47) 1.37 (0.72-2.60) NS
1 1.65 (0.4885.60) 1.10 (0.34-3.57) NS
1.32 (0.74-2.35) 1.01 (0.74-1.37) NS
1.82 (0.88-3.78) 1.44 (0.88-2.36) NS
1.61 (0.49-5.29) 0.49 (0.15-1.64) NS
1.00 (0.60-1.68) 0.74 (0.41-1.35) NS
1.51 (0.76-3.00) 1.05 (0.34-3.19) NS
0.88 (0.49-1.57) 0.96 (0.71-1.28) NS
1.11 (0.55-2.26) 1.22 (0.74-2.01) NS
1.05 (0.42-2.61) 0.21 (0.02-2.62) NS 1 0.86 (0.28-2.59) 1.21 (0.37-3.93) NS
1.09 (0.62-1.92) 0.86 (0.63-1.16) NS
1.27 (0.66-2.42) 1.01 (0.59-1.75) NS
1.38 (0.76-2.51) 1.19 (0.88-1.62) NS
1.75 (0.89-3.43) 1.83 (1.19-2.83) P= 0.0158
1.18 2.85 P 1 0.88 0.94
(0.60-2.34) (1.55-5.25) = 0.0003
1.54 (0.73-3.22) 2.35 (1.61-3.45) P = 0.0000
(0.49-1.57) (0.70-1.25) NS
0.54 (0.26-1.14) 0.90 (0.48-1.69) NS
0.97 (0.56-1.68) 0.50 (0.26-0.94) P= 0.0415
0.92 (0.50-1.70) 1.16 (0.79-1.70) NS 1 0.76 (0.36-1.62) 1.22 (0.54-2.78) NS
0.61 (0.35-1.09) 0.54 (0.30-0.98) P = 0.0373
0.55 (0.26-1.13) 0.68 (0.38-1.20) NS
0.53 (0.16-1.69) 0.92 (0.33-2.60) NS
0.62 (0.35-1.09) 0.49 (0.27-0.90) P= 0.0171
0.59 (0.31-1.13) 0.59 (0.34-1.02) NS
0.51 (0.19-1.34) 0.37 (0.14-0.94) P= 0.0346
0.79 (0.45-l .38) 0.54 (0.30-1.00) P = 0.0507
0.68 (0.32-1.44) 0.86 (0.41-1.80) NS
0.99 (0.28-3.51) 1.69 (0.44-6.43) NS
0.56 (0.31-1.01) 0.80 (0.60-1.07) NS
0.59 (0.30-1.14) 0.83 (0.53-1.32) NS 1 0.80 (0.39-1.62) 0.47 (0.22-1.00) P = 0.0486 1 0.75 (0.38-1.48) 1.31 (0.81-2.13) NS
0.54 (0.18-1.58) 0.36 (0.13-0.99) NS
0.92 (0.53-1.59) 0.43 (0.22-0.82) P= 0.0145 0.64 (0.35-1.16) 0.96 (0.70-1.24) NS
*Adjusted for energy. **Adjusted for nutrients which are significant when unadjusted or adjusted for energy. f value, for linear trend of odds ratio. NS, non-significant.
0.98 (0.38-2.54) 0.78 (0.23-2.69) NS 1 1.32 (0.55-3.15) 1.38 (0.62-3.08) NS 1 1.57 (0.62-3.97) 2.02 (1.06-3.85) P = 0.0229
0.54 (0.22-1.31) 1.56 (0.64-3.84) NS
0.47 (0.17-1.25) 0.44 (0.17-1.18) P= 0.0215 0.41 (0.14-1.16) 0.81 (0.28-2.34) NS
Diet and Prostate Table 4.
Variables
sign$cantly
related to prostate
cancer
after
adjustment for possible confounders*
Tertiles
Variable Protein
Saturated
fatty acid
Fibre
a-Tocopherol
Vitamin
B 12
Calcium
Iron
1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
*Energy, protein, fat-total, sugar-total, fibre, retinol, acid, vitamin B12, sodium, nesium and iron.
Odds ratio 1 1.71 13.54 1 1.54 3.63 1 0.83 4.02 1 1.10 0.15 1 1.74 2.07 1 0.35 0.37 1 0.40 0.34
95% confidence interval
Significance of linear trend
0.56-5.24 2.38-77.13
NS
0.60-3.95 1.03-12.79
NS
0.27-2.51 1.38-11.73
0.0564
0.45-2.74 0.05-0.53
0.0117
0.66-4.56 1.08-3.97
0.0200
0.12-1.00 0.14-0.99
0.0189
0.14-1.11 0.12-0.95
0.0089
saturated fatty acids, carbohydrate, retinol equivalent, a-tocopherol, folic potassium, calcium, phosphorus, mag-
After adjustment for energy, a significant relationship with prostate cancer was found with high protein intake, fat-total, retinol (moderate and high intake), retinol equivalent, a-tocopherol, folic acid, vitamin B12 and iron. A linear trend was significant for retinol, retinol equivalent, folic acid, vitamin B12 and iron. After adjustment for nutrient variables significantly related to prostate cancer using unadjusted or adjusted for energy, a significant relationship with prostate cancer was found for high energy intake, protein, saturated fatty acids, fiber, LXtocopherol, vitamin B12, calcium and magnesium. A linear trend was significant for fibre, a-tocopherol, vitamin B12 and calcium. In Table 4 are presented odds ratios and 95% confidence intervals for nutrients sigmficantly related to prostate cancer after adjustment for all factors significantly associated with prostate cancer according to previous analyses. High protein intake, saturated fatty acid, fibre and vitamin B12 appeared to be risk factors. A protective effect was found for a high intake of a-tocopherol, calcium and iron. When all these variables, significantly related to prostate cancer according to previous analyses, were included in the
Table 5. Risk factors for prostate cancer-multivariate
Nutrient Retinol equivalent (mg) Vitamin B 12 (mg) Iron (mg) Constant
B 0.4940 0.6258 -0.9279 -1.3905
Cancer
105
model of logistic regression step by step analysis, risk factors for prostate cancer appeared to be retinol equivalent and vitamin B12, and a protective factor was iron (Table 5). The results did not change when other non-dietary factors significantly related to prostate cancer were included in the model (occupational physical activity during the year preceding the disease; specific occupational exposure, that is, exposure to asbestos, steel, dyes and lacquers, bitumen, pitch, iron, nickel, lead, fertiliser and other potentially harmful agents; nephrolithiasis; “other” diseases such as chronic bronchitis, chronic rheumatic diseases, hypertension, cardiomyopathia, diabetes mellitus, renal diseases, eye diseases and tuberculosis; greater number (2 3) of brothers; greater number (2 8) of sexual partners [6]. The reason why retinol equivalent was not found to be related to prostate cancer when adjustment was made for vitamin B12 was the high colinearity between these two variables (Pearson coefficient of linear correlation r = 0.7636, P= 0.0000). Since some investigations suggested that the vitamin A effect depends on fat intake, we observed the relationship of prostate cancer with retinol equivalent and carotene in participants whose average fat intake was low (lower than the median value of the control group) and those whose average fat consumption was high (> median values for control group). Total fat and animal fat intake were observed separately. Retinol equivalent was directly related to prostate cancer irrespective of the quantity and type of fat consumed, although subjects whose animal fat intake was > median value had much higher relative risk in comparison with those whose animal fat intake was below median (OR= 4.68, 95% CI = 3.80-5.57 and OR = 1.90, 95% CI = 1.20-2.61, respectively). In those whose total fat intake was > median, the odds ratio for retinol equivalent was 3.04 (2.08-4.00) and in participants whose total fat intake was below the median, the odds ratio was 3.53 (2.86-4.19). For carotene intake such an association was not present. DISCUSSION The main dietary component which has been associated with prostate cancer risk is fat [l]. Ecological studies and the majority of case-control studies [7-91 have shown a strong positive association. However, cohort studies [ 1 O-l 21 have not been consistent in their findings with regard to dietary fat. Snowdon and associates [lo] found that prostate cancer mortality was positively associated with the consumption of milk, cheese, eggs and meat. In the study of Giovannucci and associates [l 11, total fat was directly related to risk of advanced prostate cancer-the association being due primarily to animal fat. In contrast, in Severson and associates’ study [12], there was no indication that heavy consumption of fat increased the risk of prostate can-
analysis (cases were divided at each tertile level of controls)
Standard 0.2484 0.2502 0.1993 0.4282
error
P value
Odds ratio
95% confidence interval
0.0467 0.0124 0.0000 0.0012
1.64 1.87 0.40
1 .Ol-2.67 1.15-3.05 0.27-0.58
106
H.D. Vlaiinac
cer, and Hsing and associates [13] also observed no association with consumption of meat, eggs or dairy products. It has been hypothesised that dietary fat affects cancer occurrence via alteration in the hormone environment. Some experimental data [ 14, 151 and analysis of hormone levels in the human population provide some corroborative information. A Western diet fed to black South African men increased the urinary excretion of oestrogen and androgens, while the excretion of oestrogen and androgens decreased in black North American men fed a vegetarian diet [16]. Bishop and associates [17] correlated differences in sex hormone levels in 171 twin pairs with differences in reported dietary habits. They found differences in fat intake correlated with differences in testosterone levels. Which fat component in particular may be responsible for the association with prostate cancer has not yet been established [l]. The association with saturated fat observed in our study had been most frequently found [ 18, 191 or suggested by the association with animal fat and with meat or dairy products [20]. However, West and associates [21] found a relationship with saturated and unsaturated fat. Giovannucci and associates [ 1 l] found that saturated fat, unsaturated fat and a-linolenic acid were associated with advanced prostate cancer risk, but only an association with cc-linolenic acid persisted when saturated fat, unsaturated fat, linolenic acid and a-linolenic acid were modelled simultaneously. Data about the relationship of vitamin A or its precursors, especially beta-carotene, to prostate cancer risk, are also inconsistent both in epidemiological and in animal studies [I]. Our findings that retinol and retinol equivalent are risk factors for prostate cancer, with an odds ratio of around 2.00, is in accordance with many case-control [22-251 studies and some cohort studies [ 12, 13, 261. Studies in which an inverse association has been found are also numerous [20, 27, 281. Methodological differences in study design and dietary survey methods might, in part, explain these inconsistencies, but opposite results were obtained even in studies using the same method for dietary data collection. The suggestion that the effect of vitamin A or its precursors depends on fat intake [28] is not supported by our data, showing a positive effect of retinol equivalent and no effect of carotene, independently of quantity and type of fat consumed. Nevertheless, it is possible that moderate amounts of vitamin A might protect against prostate cancer, while large amounts could enhance the risk [28]. According to a recent report, a trial in the United States involving 18 000 smokers has been discontinued 2 years early after initial results showed that taking vitamin A and beta-carotene as supplements may increase the risk of cancer [29]. Data on the effect of energy intake are also inconsistent. Severson and associates [ 121 did not find that total energy intake caused an increased risk of prostate cancer. According to the study of Rose and associates [30], prostate cancer mortality rates were only weakly associated with total calorie intake in the 28 countries, due to a strong positive correlation with calories of animal origin. In the study of West and associates [21], energy intake had an odds ratio of 2.5 (95% CI= 1.0-6.5). In our study the highest tertile of total energy intake was a risk factor for prostate cancer after adjustment for possible confounders including protein, tota fat, carbohydrate and total sugar.
et al.
The positive association between prostate cancer and protein intake found in our investigation was noted in studies of Kolonel and associates [31], Heshmat and associates [32] and Metlin and associates [20], but was not found in other studies [12, 271. The effect of total caloric intake as well as protein would most likely be through hormones. The positive relationship between prostate cancer and fibre intake found in this investigation would also most likely be through hormones. Lubin and associates [33] found that breast cancer risk decreases with a higher consumption of food rich in fibre, even in those with a high intake of fat and protein. Adlercreutz [34] suggested that breast cancer risk may be favourably influenced by lignans, precursors of which occur in the diet and are associated with fibre component. He hypothesised that the fibre had an anti-oestrogen effect. Since oestrogen has a beneficial effect in the treatment of prostate cancer, it is postulated that a decrease in the oestrogen level could increase the risk of prostate cancer [35]. At the same time, the lower risk of prostate cancer in cirrhotic patients could be the result of hyperoestrogenism that is present in subjects with cirrhosis [36, 371. There is no literature data on vitamin B12 as a risk factor for prostate cancer. Vitamin B12 is involved in all metabolic processes, but at present we have no adequate explanation for its positive connection with prostate cancer. The protective effect of vitamin E and iron, observed in could have a plausible explanation. the present study, Vitamin E is known as an intracellular antioxidant [38]. It also stimulates T cells and increases immunological reactions [39]. In vitro and in viva vitamin E inhibits the nitrosamine synthesis of N-nitroso compounds, and nitrosamide, which are known to be carcinogenic in laboratory animals [40]. The protective effect of vitamin E has been reported for different cancers including colon cancer, head and neck cancers, cervical cancer [41] and breast cancer [42]. A study is ongoing to investigate whether S(-tocopherol will increase the DNA repair capacity in patients at risk of head, neck and colon cancer [43]. The fact that iron is involved in oxidation-reduction processes makes it plausible that iron may decrease the risk of cancer by inhibition of free radical reactions. There are no data about the relationship of prostate cancer and calcium intake, but the protective effect of calcium has been reported for some other cancers [44, 451. Two clinical trials are ongoing to investigate the effect of calcium supplementation in colon cancer chemoprevention [46] and in preventing recurrences of neoplastic polyps of the large bowel [47]. The relationship of prostate cancer with vitamin B12, vitamin E, fibre, calcium and iron should be corroborated by other investigations.
Nomura AMY, Kolonel LN. Prostate cancer: a current perspective. Am 3 Epidemiol 1991, 13, 200-227. Jain D. Diet history: questionnaire and interview techniques used in some retrospective studies of cancer. 3 Am Diet Assoc 1989,89, 1647-1652. Kaic-Rak A, Antonic K. Tables of Food Items Composition. Zagreb, Public Health of Croatia, 1990. Breslow NE, Day NE. Statistical Methods in Cancer Research. Volume I-The Analysis of Case-Control Studies. Lyon, IARC, 1980, 248-279.
Diet
and
Prostate
5. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York, Wiley-Interscience 1989, 82-134; 187-215. 6. IliC M, Vlajinac H, Marinkovic J. Case-control study of risk factors for prostate cancer. Br J Cancer 1996, 74, 1682-1685. 7. Berg JW. Can nutrition explain the pattern of international epidemiology of hormonal-dependent cancers? Cancer Res 1975,35,3345-3350. 8. Kolonel LN, Hankin JH, Lee J, Chu SY, Nomura AMY, Hinds MW. Nutrient intakes in relation to cancer incidence in Hawaii. Br3 Cancer 198 1, 44, 332-339. 9. Slattery ML, Schumacher MC, West DW, Robison LM, French TK. Food-consumption trends between adolescent and adult years and subsequent risk of prostate cancer. Am 3 Clin Nua 1990, 52, 752-757. 10. Snowdon DA, Phillips RL, Choi W. Diet, obesity, and risk of fatal prostate cancer. Am 3 Epidemiol 1984, 120, 244-250. 11. Giovannucci E, Rimm :EB, Colditz GA, et al. A prospective study of dietary fat and risk of prostate cancer. 3 Nat1 Cancer znst 1993,85, 1571-1570. 12. Severson RK, Nomura AMY, Grove JS, Stemmennann GN. A prospective study of demographics, diet, and prostate cancer among men of Japanese ancestry in Hawaii. Cancer Res 1989,49, 1857-1860. 13. Hsing AW, McLaughlin JK, Schuman LM, et al. Diet, tobacco use and fatal prostate cancer: results from the Lutheran brotherhood cohort study. Cancer Res 1990, 50, 6836-6840. 14. Pollard M, Luckert PH. Promotional effects of testosterone and dietary fat on prostate carcinogenesis in genetically susceptible rat. Prostate 1985, 6, l-5. 15. Carroll KK, Noble RL. Dietary fat in relation to hormonal induction of mammary and prostatic carcinoma in Nb rats. Carcinogenesis 1987, 8, 85-853. 16. Hill P, Garbaczewski L, Helman I’, Walker ARE, Games H, Wynder EL. Environmental factors and breast and prostatic cancer. Cancer Res 1981, 41, 3817-3818. 17. Bishop DT, Meikie AW, Slattery ML, Stringham JD, Ford MH, West DW. The effect of nutritional factors on sex hormone levels in male twins. Gen Epidemiol 1988, 5, 43-59. 18. Kolonel LN, Yoshizawa CN, Hankin JH. Diet and prostatic cancer: a case-control study in Hawaii. Am 3 Epidemiol 1988, 127, 999-1012. 19. Hankin JH, Zhao Ll’, Wilkens LR, Kolonel LN. Attributable risk of breast, prostate and lung cancer in Hawaii due to saturated fat. Cancer Causes Control 1992, 3, 17-23. 20. Mettlin C, Selenskas S, Natarajan N, Huben R. Beta-carotene and animal fats and their relationship to prostate cancer risk. A case-control study. Cancer 1989, 64, 605-612. 21. West DW, Slattery ML, Robinson LM, French TK, Mahoney AW. Adult dietary intake and prostate cancer risk in Utah: a case-control study with special emphasis on aggressive tumors. Cancer Causes Control 19511, 2, 85-94. 22. Graham S, Haughey B, Marshall J, et al. Diet in the epidemiology of carcinoma of the prostate gland. 3 NatZ Cancer Inst 1983, 70, 687-692. 23. Heshmat MY, Kani L, Kovi J, et al. Nutrition and prostate cancer: a case-control study. Prostate 1985, 6, 7-17. 24. Kolonel LN, Hankin JH, Yoshizawa CN. Vitamin A and prostate cancer in elderly man: enhancement of risk. Cancer Res 1987,47, 2982-2985. 25. Talamini R, La Vecchia C, Decarli A, Negri E, Franceschi S. Nutrition, social factors, and prostatic cancer in a Northern Italian population. Br 3 Cancer 1986, 55, 8 17-82 1. 26. Paganini-Hill A, Chao A, Ross RK, Henderson BE. Vitamin A, beta carotene and the risk of cancer: a prospective study. 3 NatZ Cancer Inst 1987, 79, 443-448. 27. Ross RK, Shimicu H, Paganini-Hill A. Case-control studies of prostate cancer in blacks and whites in southern California. 3 NatZ Cancer Inst 1987, 78, 869-874.
28.
29. 30.
3 1.
32. 33.
34.
35. 36. 37.
38. 39.
40.
41.
42.
43.
44. 45.
46.
47.
Cancer
107
Ohno Y, Yoshida 0, Oishi K, Okada K, Yamabe H, Schoeder FH. Dietary beta carotene and cancer of the prostate: a casecontrol study in Kyoto, Japan. Cancer Res 1988, 48, 13311336. Cookson C. Cancer-risk fears halt US tests on vitamin pills for smokers. Financial Times 1996, January 20121, 24. Rose DP, Boyar AI’, Wynder EL. International comparisons of mortality rates for cancer of the breast, ovary, prostate and colon, and per capita food consumption. Cancer 1986, 58, 2363-237 1. Kolonel LN, Hankin JH, Lee J, Chu SY, Nomura AMY, Ward Hinds M. Nutrient intakes in relation to cancer incidence in Hawaii. Br3 Cancer 1981, 44, 332-339. Heshmat MY, Kani L, Kovi J, et al. Nutrition and prostate cancer: a case-control study. Prostate 1985, 6, 7-17. Lubin F, Wax Y, Modan B. Role of fat, animal protein and dietary fiber in breast cancer etiology. A case-control study. 3 Nat1 Cancer Inst 1986, 77, 605-612. Adlercreutz H. Does fiber-rich food containing animal lignan precursors protect against both colon and breast cancer? An extension of the “fiber hypothesis”. Gasroenterology 1984, 86, 761-766. Resnik M. Hormonal therapy in prostatic carcinoma. UroZogy 1984, 24, 18-20. Glantz MG. Cirrhosis and carcinoma of the prostate gland. 3 Ural 1964,91, 291-293. Chopra IJ, Tulchinsky D, Greenway FL. Estrogen-androgen imbalance in hepatic cirrhosis. Ann Intern Med 1973, 79, 19% 203. Bieri JG, Corash L, Hubbard VS. Medical use of vitamin E. N EngZJ Med 1983, 308, 1063-1071. Raletic-Dragicevic J, Djulizibaric J. Can vitamins have preventive effect on cancer? Information About Cuncerogenics (in Serbian) 1985, 2, 10-12. Choi BCK. N-nitroso compounds and human cancer: a molecular epidemiologic approach. Am 3 Epidemiol 1985, 121, 737-743. Verreault R, Chu J, Mandelson M, Shy K. A case-control study of diet and invasive cervical cancer. IntJ Cancer 1989,43, 1050-1054. Schrauzer GN, White DA, Schneider SJ. Cancer mortality correlation studies. III Statistical association with dietary selenium intakes. Bioinorg Chem 1977, 7, 23-34. Berwick M, Schantz SP, Liebes L, Shike M. Vitamin E and DNA repair in patients at risk for head and neck cancer and colon cancer. In Sankaranaryanan R, Wahrendorf J, Demaret E in collaboration with Becker N, eds. Directoy of On-Going Research in Cancer Epidemiology. Lyon, IARC, 1996, 463. Rosen M, Nystrom L, Wall S. Diet and cancer mortality in the counties of Sweden. Am 3 Epidemiol 1988, 127, 42-49. Meyer F, White E. Alcohol and nutrients in relation to colon cancer in middle-aged adults. Am 3 Epidemiol 1993, 138, 225226. Wargovich MJ, Steinbach G, Winn R, Lee J, Levin B. Calcium and aspirin prevention trial. In Sankaranaryanan R, Wahrendorf J, Demaret E in collaboration with Becker N, eds. Directoy of On-Going Research in Cancer Epidemiology. Lyon, IARC, 1996, 437. Baron JA, Rothstein R, Beck G, er al. Calcium in the prevention of neoplastic polyps. In Sankaranaryanan R, Wahrendorf J, Demaret E in collaboration with Becker N, eds. Directoy of On-Going Research in Cancer Epidemiology. Lyon, IARC, 1996, 429.
Acknowledgement-Funded Technology of Serbia through
by the Ministry contract No. 8774,
for Science 1991-5.
and